DESCRIPTION: (Applicant's Description) Among several hundred known cancer-associated genes, a substantial number remain to be clearly assigned to biochemical pathways. In addition, as more efficient genomic methods are developed, the number of potential cancer genes is increasing rapidly. It is therefore important to develop and exploit high-throughput methods to reveal functional interactions among components of signaling pathways. We describe our proposed contribution to assign key breast gene products to particular signaling pathways and determine the steps at which they function. The strategy is based on the identification and validation, using the two-hybrid system, of potential physical interactions mediated by gene products related to normal breast development and/or breast tumor formation. We have recently developed a version of the two-hybrid system that solves most of the common problems of the methods, including the relatively high number of false positives In addition, since the current versions of the two-hybrid system also fail to provide a clear link between potential physical interaction in yeast and functional interaction in vivo, we have developed a "reverse two-hybrid" system that efficiently generates alleles that are defective for protein:protein interaction or trans-acting dissociating molecules. Such reagents can be used to assess the functional significance of potential interactions in the relevant in vivo biological system. This section of the Program Project Grant describes a Functional Genomics Core referred to as "Protein Interaction Core" or "Scientific Core B". The Protein Interaction Core is composed in turn of two components: 1) A "Service Component" aimed at performing forward and reverse two-hybrid screens for the other projects described in the Program, and 2) A "Technology Development Component" aimed at developing functional genomic strategies to identify interacting proteins for the currently known key breast gene products. We plan to develop a semi-automated version of the two-hybrid system to identify protein:protein interactions in large-scale screens. In addition, we plan to develop semi-automated version of the reverse two-hybrid system which should provide functional information about the protein:protein interactions identified. The resulting information will be released in the form of a publicly available database, providing a tool for the community of breast cancer researchers, including our colleagues on this grant.